Estimating causal effects with non-experimental data
Embargo until
Date
2017-07-28
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
In this manuscript we seek to relax some of the traditional assumptions associated with the estimation of causal effects. In particular, we relax the assumption that all confounders are measured without error and the assumption that the observations in the sample are independent and identically distributed. Furthermore, we explore the impact of model misspecification in the estimation of population causal effects.
Description
Keywords
Non-experimental data, Causal Inference, Survey Weights, Complex Survey Data, Measurement Error, Model Misspecification, ATT, PATT, SATT, ATE, PATE, SATE